Monday, April 30, 2012

Modernizing SAS BI

Ajay Ohri at DecisionStats posted an article recently about a software product which claims to be able to convert the SAS statistical processing language into Java. If so, that could mean that SAS customers could see a tremendous savings in licensing fees.

"Profound" is how Dulles Research, the vendor of this tool called Carolina, describes the cost savings of Java over SAS. Serial entrepreneur Drew Eginton owns the company. 

While automatically translating SAS into Java may be technically possible, a conversion undertaking of this magnitude would still involve a serious amount of parallel testing. 

Companies will also run into plenty of resistance from their in-house SAS experts. In fact, they should expect their mainframe capacity planners to revolt. These guys traditionally use the MXG product from Barry Merrill to calculate the system usage. Dulles Research has evidently already run into this bunch, as they put this statement on their main webpage: "Run MXG SAS-free." 

Statistical PROCs
Of course, SAS is full of statistical functions so Dulles Research will need to have a Java replica for each:
  • PROC ANOVA
  • PROC CANCORR
  • PROC CANDISC
  • PROC CLUSTER
  • PROC CORR
  • PROC DISCRIM
  • PROC FACTOR
  • PROC FASTCLUS
  • PROC FREQ 
  • PROC FUNCAT
  • PROC GLM
  • PROC MATRIX
  • PROC MEANS
  • PROC NEIGHBOR
  • PROC NESTED
  • PROC NLIN
  • PROC NPAR1WAY
  • PROC PLAN
  • PROC PRINCOMP
  • PROC PROBIT
  • PROC RSQUARE
  • PROC RSREG
  • PROC SCORE
  • PROC STANDARD
  • PROC STEPDISC
  • PROC STEPWISE
  • PROC TABULATE 
  • PROC TREE
  • PROC TTEST
  • PROC VARCLUS
  • PROC VARCOMP

More PROCs
SAS has plenty of "general" procedures as well. Any translator would have to handle:
  • PROC APPEND
  • PROC BMDP
  • PROC BROWSE
  • PROC CALENDAR
  • PROC CHART
  • PROC CONTENTS
  • PROC CONVERT
  • PROC COPY
  • PROC DATASETS
  • PROC DELETE
  • PROC EDITOR
  • PROC FORMS
  • PROC GPLOT
  • PROC PDS
  • PROD PDSCOPY
  • PROC PLOT
  • PROC RELEASE
  • PROC REPORT 
  • PROC SOURCE
  • PROC TAPECOPY
  • PROC TAPELABEL
  • PROC TEMPLATE 
  • PROC TRANSPOSE

Even More Things to Convert
The Java replica would have to handle all of the basic file input and output handling as well, such as:
  • Basic CALL SYMPUT
  • DATA/CARDS 
  • DATA/INPUT statement
  • DATA/MERGE
  • DATA/UPDATE
  • EXPORT statement
  • Macro definition and invoking
  • ODS CSV, RTF, PDF, PS
  • ODS HTML
  • ODS PRINTER
  • OS Commands
  • PROC IMPORT
  • PROC RANK basic
  • PROC SUMMARY basic
  • PROC UNIVARIATE
  • STYLE=
  • Traffic-Lighting

SAS also has its own data structure, so eliminating SAS would mean converting any existing data into another format. The replacement language would need to be able to access that new format.

When I investigated translating SAS to the WebFOCUS BI product (based on a 4GL programming language), I personally considered some of the following SAS features as stumbling blocks:
  • Symbolic variable processing
  • DATA/INPUT options
  • DATA/SET multiple
  • ODS OUTPUT
  • ODS SELECT/EXCLUDE
  • ODS TRACE
  • OUTPUT statement for multiple output files or DO/END loops
  • Procedural scripting
  • Selection by observation number
  • Summary set processing
  • Temporary variables as arrays

For example, the WebFOCUS 4GL language has no construct for arrays; there just isn't anything like that. So figuring out how to convert SAS array processing into WebFOCUS array processing is something I would have to think hard about to solve.

Of course, Java is a lower-level language than the WebFOCUS 4GL so Dulles Research could write their own classes to replicate anything that SAS can do. It might take a while, but it is definitely possible. 

While SAS is one of the hottest BI technologies on the market, its licenses can be very expensive. Many companies, especially those with mainframe licenses, are looking for ways to reduce that cost. 

If you are considering replacing SAS with some other product, contact me about our rapid assessment engagement using the BI Analyzer product to scan your SAS applications (include supplemental technologies such as FOCUS, JCL, CLists, etc.) and save the results into a data repository from which we can quickly run analytics on the size of the modernization effort.


You may also be interested in these articles:


Wednesday, April 25, 2012

Down and Klout in the Magic Kingdom

Online influence tracker Klout announced today the release of their official iPhone app. See their blog for more information or download it from the App Store.

If you have never checked your Klout score, you should do it now. A recent Wired article explains that your next job interviewer may ask for your score and that some companies may be targeting you (or ignoring you) as a customer based on your score. 

The people at Klout explain it this way:
The Klout Score measures influence based on your ability to drive action. Every time you create content or engage you influence others. The Klout Score uses data from social networks in order to measure: 
  • True Reach: How many people you influence
  • Amplification: How much you influence them
  • Network Impact: The influence of your network


Early in 2011, Robert Paterson talked to Joe Fernandez, the CEO of Klout, about the new online world of influence. Of course, author Cory Doctorow saw this coming a decade ago. 

Doctorow's book, "Down and Out in the Magic Kingdom," is about a 22nd-century American society where there is neither death nor money. Democracy is gone and the prevailing mode of government is "adhocracy." Instead of monetary currency, people use online influence credits; the more well-known you are, the "richer" you are (shades of Klout!). 

While you wait for your Klout score to increase, read the relevant sci-fi book (like Klout, it's free!). And, by the way, if you like this information be sure to give me a K+. 

Sunday, April 22, 2012

Gartner says Predictive Analytics is the Hot BI Topic

Rita Sallam, BI analyst for Gartner, says that the use of predictive analytics software will be pervasive by 2020. By pervasive, she means that three-quarters of all users will have access to predictive functionality. Today, that figure is less than one-third, while Sallam expects it to grow to half of the users in the next two years.
“Analytics is the combustion engine of business, and it will be necessary for organizations that want to grow, innovate and optimize efficiency. Given its far-reaching impact, it is one of the few software markets that thrive even in adversity.”

Interestingly, Rita points out that the idea of selecting a BI standard stack based on a particular mega-vendor may be a thing of the past, especially when it comes to predictive analytics software.
“A stack-centric mentality and single vendor standardization policies won't cut it. Understand that your organization needs a portfolio of analytic capabilities.”

For more information, see this article from Enterprise Apps Today.

Saturday, April 21, 2012

Micro Focus Partners

While I learned lots of things at this week's Micro Focus 2012 Partner Conference in Dallas, Texas, one insight stuck out in my mind.

Micro Focus sees themselves, their products, and their market very clearly.

Clear Software Business Model 
They understand themselves to be a software vendor. That means not trying to be a services vendor even if it could result in additional revenue.

Only about six percent of their total revenue comes from consulting services. In fact, Micro Focus has intentionally scaled back Professional Services group to be just high-level architects and consultants. They give the remaining business (perhaps quite lucrative) to services partners, even the prime contracts to lead an entire initiative.

Micro Focus's partners range from giants (think Microsoft, Accenture, Dell, HP, and Infosys) to small boutique firms (think Partner Intelligence).

In 2004, MIT Professor Michael Cusumano wrote on the topic of software vendor business models in his book, "The Business of Software," saying that running a software firm and services firm differs greatly.

Software vendors concentrate on making and selling product, with an important financial metric being a high ratio of revenue per headcount (make as much money as possible with as few people).

For example, Micro Focus is a highly profitable software vendor with 1200 employees and annual revenue of about $450 million. Doing the math, that would come up to about $375K per person--an amazing ratio for any software vendor.

Cusamano wrote:
"...if it costs you roughly the same to make one copy or one million copies of a product, you would be a fool not to want to make and sell a million copies of every product you create."

On the other hand, services vendors perform labor-intensive work. Instead of focusing on selling lots of software product, services companies need lots of people and must maximize billing and utilization rates. These companies can only make as much money as they have people and hours in the day. To match the Micro Focus revenue per headcount, a services vendor would have to keep every employee fully utilized at an hourly bill rate of about $210.

Cusamano admits it is not impossible to be a "hybrid" vendor offering both products and services:
"One reason why software companies, even those with products to sell, may have high services revenues is that their technologies are often too complex to package as 'off-the-shelve' products. As a result, they sell 'solutions' that require customizations (say, 20 to 50 percent of the total amount of code) or special integration and installation work."

Micro Focus will concentrate on the software product while specialized services partners provide the rest of the "ecosystem" of the total solution.  Micro Focus acknowledges that the market is too large and complex for one vendor to attempt by itself.

Of course, Micro Focus's partners do not specialize in just consulting services. One executive said that Micro Focus has about 350 relationships within their three solution areas of Mainframe technology, Distributed COBOL products, and Borland products. Many of these partners are ISVs who embed Micro Focus products into third-party software applications.

Clear Product Positioning 
In addition to understanding their nature as a pure software vendor, Micro Focus also sees their product lines clearly. At one point in the Q&A session, the Micro Focus facilitator stopped the person in the audience asking a question and said, "Let me clarify, we don't do COBOL anymore. We do Visual COBOL."

This is an important point. Micro Focus understands that the market for legacy mainframe COBOL in business systems is, for all intensive purposes, gone. COBOL, of course, is the "crown jewels" for Micro Focus. That means that they are concentrating on the large "contemporary, distributed" market for Windows-based visual IDEs such as Eclipse and Microsoft's Visual Studio utilizing managed frameworks such as .NET and Java.

One such managed language that works with either .NET or Java frameworks within GUI development tools is COBOL. Micro Focus must now work hard to change the perception of COBOL as being an old language so that it gains more acceptance in the distributed application development market.

They are also working hard to provide alternatives to the mainframe platform. Micro Focus offers an entire "stack" of products:

  • Enterprise Analyzer (for visually analyzing the applications)
  • Enterprise Developer (Personal and Team Editions)
  • Enterprise Test Server (for moving app dev workload off of the tradition mainframe platform)
  • Enterprise Server (the flagship mainframe-alternative engine)

However, Micro Focus is not taking an extreme position of "just eliminate the mainframe." Rather than be a complete adversary to IBM, they can offer complementary solutions. For example, their Enterprise Test Server platform allows a mainframe customer to keep production applications and data running on the mainframe and just move application development and testing CPU cycles to a cheaper platform that mimics the mainframe.

Transparency 
I would not have gained this insight into Micro Focus had they not been so transparent with their partners about their business. Executives openly discussed past issues, changes in positioning, and future plans during the conference. They demonstrated tools for the partners and facilitated interactive discussions on how to best enable this partner channel through which they would provide the best services possible to their customers. They offered to provide their Sales Kickoff documentation to partners at the same time it is rolled out internally.

Frankly, I think this level of partnering with software vendors is atypical in the marketplace.

Real collaborative relationships would be impossible if Micro Focus did not see themselves first and foremost as a pure software vendor. They understand they are in the business to make and sell software products. If Micro Focus muddied the business-model waters and also tried to gain revenue from consulting services, then third-party services firms would become evil competitors instead of close allies.

Other Software Vendors, Take Notice 
Micro Focus is a shining example to other vendors on how to build a partner ecosystem where the software firm works together with external companies to sell and deliver a complete technical solution in order to thrill customers.

Note: as an example of how competing with your partners is a bad idea, see this article on Oracle's poor performance trying to sell its Sun hardware. 

Tuesday, April 17, 2012

Free eBook on Balsamiq User Interfaces

I am a big fan of Balsamiq, the vendor of low-fidelity wireframe mockup software (see my earlier blog article).

Balsamiq recently created an eBook of interviews and have made it available free on their website. Download a copy (PDF, EPUB, or MOBI) at this website link

Monday, April 16, 2012

5 Hot BI Apps for the Apple iPad

A recent Network World article listed twenty Apple iPad apps that every CIO will want. Of those, a quarter of the mobile apps dealt with Business Intelligence, a hot topic on every business executive's mind.

Network World showcased the following BI apps for iPad tablets:


Missing from their list were the following iPad BI applications that are worthy of consideration:


See the article for more information. 

20 Years of BI

The older we get, the faster time goes by. Experts say this is because we respond to novel events while young but once we get older, things seem more routine and less worthy of attention. Our resulting lack of focus causes our sense of time to get out of whack.

As a personal example, I just realized I am approaching the twentieth anniversary of my very first presentation at Information Builders' annual user conference.

Back in 1992, I spoke on application efficiencies to a few hundred 4GL developers. That was in Anaheim, CA, near Disneyland. I spoke on  common mistakes that caused slower application responses.

Now it is 2012 and I am speaking on avoiding BI blunders in Orlando, FL, near Disney World. In particular, I will talk about the common mistakes companies make during BI dashboard software projects.

In a sense, I am still doing the same old thing twenty years later. Yet plenty is different.

For example, this thing called the Internet brought the thin-client web user interface. Cheap telecommunications showed up to give us a low-cost global workforce just in time for Y2K projects. Computers got smaller yet faster and cheaper (in fact, people can now easily carry tiny computers around all day long). Companies accumulated gigantic piles of data about their customers which have become just as valuable as the original business transactions. Firms have fewer people doing much more work.

Our real problem as adults may be that we don't interact with novelty the way we did as children. Back then, when we found something new, it was exciting to us and we took time to play with it. But now as adults, we don't bother to slow down to intentionally reflect on the impact of this new big thing. As a result, we fail to take action and can only become bystanders of the novelty rather than active participants.

I would love to stop and play with new technology. For example, why not build a mobile Apple iPad app using Objective-C? That would be fun! Well, my adult mind can give me a dozen reasons against the idea, most of which involve a lack of time caused by other responsibilities (for example, I should be creating a BI dashboard presentation to hit a print deadline of May 1st).

In reality, the length of twenty-four hours has not gotten shorter since we were young. Instead, we have just allowed too much old adult stuff to fill up the day. For me, it's time to slow down and do something novel, even if it seems childish.

Saturday, April 7, 2012

Dell Enters Application Modernization Market

Last week, Dell announced that it was acquiring Clerity Solutions, the Chicago-based global vendor specializing in migrating applications from the mainframe.

Dell also announced they would purchase Wyse Technology, a cloud client computing firm, as well as Make Technologies, "a leading global provider of application modernization software and services that reduce the cost, risk and time required to re-engineer applications."

Dell's coordinated purchase of Clerity, Wyse, and Make Technologies will position them to help accelerate the demise of the mainframe computer, moving expensive workload "onto more modern architectures, including the cloud."

Dell's competitor HP has already gone down this path.

Back in 2008, HP spent about $13 billion to acquire services heavyweight EDS. They also partnered with Micro Focus for application modernization software. Emerging out of this was an application modernization group focused on moving workload off of the mainframe. See an earlier blog post for more information.

Now, Dell and HP are in a head-to-head struggle to see who can more quickly move more legacy apps off of the mainframe onto their platforms. This is a battle to gain mid-range platform market share by stealing it from the IBM mainframe.

Should IBM be worried that two mid-range competitors are consolidating on different flanks, both intent on eliminating the mainframe computer?

Many companies want to eliminate the high cost of running legacy applications on mainframe computers. The initial big ticket price was one thing, but most software vendors annually hit their customers with 20% maintenance fees. If a decade ago you spent $1 million for a mainframe software product, you probably pay $200K each year just to keep it updated.

Over those ten years, you may have already spent $3 million on that one package; a product for which today you cannot find technical support resources. A product for which you will continue to pay annual maintenance fees despite declining usage.

Mainframe software customers got the short end of the stick, but what could they do? Their challenge has always been the extremely high cost and risk of trying to migrate off the mainframe.

Now that Dell and HP are making it easier for customers to part ways with IBM, more companies will stop just considering leaving the mainframe and try it.  

If you have a legacy 4GL (such as FOCUS, NOMAD, or RAMIS) still running on the mainframe, click here to read my 2008 blog article on the topic of migrating and modernizing them. 

Wednesday, April 4, 2012

SAP Enters Predictive Analytics Space

Software mega-vendor SAP says that it is responding to customer demand for predictive analytics software with a new product called (drum roll...) Predictive Analysis (wow, what a let down...).

While SAS is the big statistical player in this BI market, IBM entered several years ago by acquiring software vendor SPSS. Of course, Oracle also wants to be the leader.

The emerging open-source solution is R Stat, which is used by Information Builders in their WebFOCUS product.

SAP says there are five things you need to understand about predictive analytics:

1. Predictive analytics is nothing new but the emerging "big data" trend changes things.
2. While it's a "data statistician game," many business users have a need and want tools.
3. This will sit on top of their in-memory HANA which promises real-time analysis from big data.
4. Not that SAP will ignore other databases, mind you...
5. And, for the time being, this is only available via the "Ramp-Up Program" for early-release testers.


For the informal version, see the SAP blog by Mani Gill, VP/GM of SAP BI.

For the entire story, see the SAP Press Release

About Me

My Photo
Helping companies make better decisions via Business Intelligence. INTP working on the E&J. Traveler, reader, family guy, coffee drinker.